Isolated Building Pollutant Dispersion (CEDVAL A1-5)

Why this case matters

Pollutant dispersion around buildings is one of the core deliverables of computational wind engineering: the concentration of a tracer released near the ground tells you where vehicle exhaust, an accidental release or a stack plume ends up at pedestrian level. Getting it right requires three things working together that the simpler scalar cases test only in isolation: a tracer source on the geometry, impermeable building faces that the tracer cannot pass through, and a turbulent mixing model that spreads the plume realistically. This case brings all three onto a real wind-tunnel benchmark.

The CEDVAL A1-5 dataset (Environmental Wind Tunnel Laboratory, University of Hamburg, Environmental Wind Tunnel Laboratory (EWTL) and Leitl[1]) measures the concentration field around an isolated rectangular building with a near-ground source in a neutral atmospheric boundary layer. It is a reference case in the urban-dispersion model evaluation framework of COST Action 732 [2] and VDI 3783 Part 9 [3], which prescribe the validation metrics (Chang and Hanna[4]). Passing it validates the voxel scalar source, the impermeable zero-flux scalar faces and the LES turbulent-Schmidt coupling on a documented, quality-controlled dataset, isolated from the additional complexity of the street-canyon case (Street-Canyon Pollutant Dispersion (CODASC)).

Physical description

An isolated rectangular building stands in a neutral atmospheric boundary layer. The approach wind separates over the building, reattaches on the roof and forms a recirculating wake on the leeward side. A passive tracer (modelling an underground-garage exhaust) is released from a near-ground source at the base of the leeward wall, and its concentration is measured throughout the wake and at pedestrian level.

In nassu the building is voxelized: a surface band carries the no-slip velocity condition and a zero-flux scalar condition (the building faces are impermeable to the tracer). The source is a separate near-ground patch of the band that holds a fixed scalar value, emitting the tracer into the wake. The atmospheric boundary layer is fed at the inlet by the Synthetic Eddy Method (SEM), the same inlet used in the atmospheric and wall-mounted-cube cases. The subgrid mixing of the tracer is controlled by the turbulent Schmidt number \(\mathrm{Sc}_t\) (see LES Subgrid Coupling).

Geometry and scaling

The building follows the CEDVAL A1-5 model at scale 1:200:

Quantity	Value
Building (along-wind \(\times\) cross-wind \(\times\) height)	\(100 \times 150 \times 125\) mm (model), \(20 \times 30 \times 25\) m (full scale)
Lattice cell size	2.5 mm/cell (building height \(H = 50\) cells)
Building in lattice cells	\(40 \times 60 \times 50\)
Domain (lattice cells)	\(752 \times 504 \times 304\) (\(\approx 15\,H \times 10\,H \times 6\,H\))
Building position	centred \(5\,H\) from the inlet, on the domain mid-width

Governing equations

The tracer is a passive scalar transported by the resolved flow with molecular and subgrid diffusion (see Scalar Transport). The validation quantity is the dimensionless concentration:

(1)\[c^* = \frac{C\, U_{\text{ref}}\, H^2}{Q}\]

where \(C\) is the measured concentration, \(U_{\text{ref}}\) the reference velocity, \(H\) the building height and \(Q\) the source emission rate. The normalisation collapses the concentration field onto a geometry-and-source-rate-independent value that can be compared directly against the wind-tunnel data.

Simulation setup

Parameter	Value
Reynolds number on \(H\)	\(\approx 1.5 \times 10^4\)
Molecular Schmidt number \(\mathrm{Sc} = \nu / D\)	1
Turbulent Schmidt number \(\mathrm{Sc}_t\)	0.7 (CWE default)
Reference velocity \(U_{\text{ref}}\) (lattice units)	0.06 (top-of-profile)
Fluid velocity set / operator	D3Q27 / RRBGK
Fluid relaxation time \(\tau\)	0.50045 (\(\nu = 1.5 \times 10^{-4}\))
Scalar velocity set / operator	D3Q7 / RRBGK
Scalar diffusivity \(D\) (lattice units)	\(1.5 \times 10^{-4}\) (\(= \nu\), since \(\mathrm{Sc} = 1\))
LES model	Smagorinsky (\(C_S = 0.17\))
Inlet turbulence	SEM, neutral ABL log-law profile (category-2 CSV)
Building wall treatment	voxelized band: RegularizedHWBB no-slip plus ScalarHWBB zero-flux
Source treatment	near-ground leeward-base voxel patch, ScalarRegularizedDirichlet \(\phi_w = 1\)
Ground	IBM terrain with the equilibrium log-law wall model
Refinement	static level-1 and level-2 slabs around the building and wake

Important

The physical emission rate \(Q\) and the physical reference velocity \(U_{\text{ref}}\) (and its reference height) are placeholders in the current setup, pending the password-protected EWTL dataset. The lattice run uses a unit source wall value and recovers the emission rate in the post-processing step from the integrated wall flux; the absolute \(c^*\) normalisation in (1) is fixed only once the dataset values are confirmed.

Note

The top-of-profile reference velocity (\(U_{\text{ref}} = 0.06\) lattice units) sits at \(\mathrm{Ma} \approx 0.10\), at the edge of the weakly-compressible guideline \(\mathrm{Ma} < 0.1\). The velocity around the building (the region of interest) stays well below this, so the weak-compressibility assumption holds where it matters.

Reference and acceptance

The reference is the CEDVAL A1-5 wind-tunnel dataset Environmental Wind Tunnel Laboratory (EWTL) and Leitl[1], evaluated with the COST Action 732 [2] and VDI 3783 Part 9 [3] metrics for urban dispersion models, following the recommended scoring of Chang and Hanna[4]. The acceptance targets are the standard quality bands:

• factor-of-two of observations \(\mathrm{FAC2} \geq 0.5\),

• normalised mean-square error \(\mathrm{NMSE} < 4\),

• fractional bias \(\mathrm{FB}\) within \([-0.3, 0.3]\),

• hit rate \(q \geq 0.66\) (with tolerance \(D = 0.25\)).

Results

Note

The quantitative comparison notebook (the dimensionless concentration \(c^*\) against CEDVAL A1-5 and the COST 732 / VDI 3783-9 metrics) will be added once the GPU validation runs are committed and the EWTL dataset values for \(Q\) and \(U_{\text{ref}}\) are confirmed, mirroring the workflow of the other v2.0 scalar cases.

---

[1] (1,2)

Environmental Wind Tunnel Laboratory (EWTL) and Bernd Leitl. CEDVAL: compilation of experimental data for validation of microscale dispersion models. Database, Meteorological Institute, University of Hamburg. Accessed via the EWTL CEDVAL database. URL: https://www.mi.uni-hamburg.de/en/arbeitsgruppen/windkanallabor/data-sets.html.

[2] (1,2)

Michael Schatzmann, Helge Olesen, and Jorg Franke, editors. COST 732 Model Evaluation Case Studies: Approach and Results. COST Office, Hamburg, 2010.

[3] (1,2)

Verein Deutscher Ingenieure. VDI 3783 part 9: environmental meteorology — prognostic microscale wind field models — evaluation for flow around buildings and obstacles. Technical Report VDI 3783 Part 9, Beuth Verlag, Berlin, 2005.

[4] (1,2)

Joseph C. Chang and Steven R. Hanna. Air quality model performance evaluation. Meteorology and Atmospheric Physics, 87(1-3):167–196, 2004. doi:10.1007/s00703-003-0070-7.